U‑Capture and MS Teams
This topic provides general architecture details for the Uniphore MS Teams integration. For system deployment details specific to a customer environment please refer to the customer's U-Capture Solution Design Document from Uniphore.
Integration Overview
Uniphore provides a realtime (active) recording solution for Microsoft Teams. In a U‑Capture / MS Teams integration, calls from MS Teams (both internal and external) and PSTN calls (with and without direct routing) are captured. The architecture diagram encompasses these call types under the “telephony” component.
When a call is initiated, MS Teams users are notified on screen stating that the call is being recorded – “non-Teams” users will also receive an audio announcement to the same effect. The Uniphore Compliance Bot captures call audio and metadata in realtime and sends it to the U‑Capture Collector. This integration type supports access to live data and post-call data via the U‑Capture App and U‑Capture APIs.
Teams Multi-Tenancy
The U‑Capture / MS Teams integration supports Teams multi-tenancy with call audio and call metadata from multiple instances of MS Teams.
In a realtime multi-tenancy model, each MS Teams tenant has call audio and call metadata captured by a single Compliance Bot, or pair of Compliance Bots in a resilient environment. The Compliance Bot feeds call audio and call metadata directly to the U‑Capture Collector before the Collector’s built-in multi-tenancy capability splits the recorded users into business units.
Resiliency
Active/Active Resiliency
An Active/Active (2N, Parallel) resiliency model has two Compliance Bots that simultaneously capture the same call audio and call metadata. Each Compliance Bot sends the call audio and call metadata to its associated Primary Collector in realtime, both of the identical calls will be present in U-Capture.
If one Compliance Bot fails for any reason, the other Bot simply continues to send its data to its Collector with no loss of recorded data or interruption of recording services. Once the failed Bot is brought back online, parallel data capture will resume.
If one Collector fails for any reason, the other Collector simply continues to send its data to the Core Services with no loss of recorded data or interruption of recording services. Once the failed Collector is brought back online by Docker, parallel data capture will resume.
Active/Standby Resiliency
An Active/Standby (N+1, Failover) resiliency model has two Compliance Bots that simultaneously capture the same call audio and call metadata. Each Compliance Bot sends the call audio and call metadata to its associated Collector in realtime.
If the Primary Collector and/or Bot fails, the Standby Collector and Bot pair will be brought online and start data capture as the new Active Collector/Bot. Any recordings that were ongoing at the time of failure will be “interrupted”. Once the failed pair is recovered, it will become the new Standby Collector/Bot.
